CHAPTER ONE

Introduction

As the nation’s economic activities, security concerns, and stewardship of natural resources become increasingly complex and globally interrelated, they become ever more sensitive to environmental conditions. For the past several decades, forecasts of weather, ocean, and other environmental phenomena made a few days ahead have yielded invaluable information to improve decision-making across all sectors of society (Lazo et al., 2011). Enhancing the capability to forecast environmental conditions outside the well-developed weather timescale—for example, extending predictions out to several weeks and months in advance—could dramatically increase the societal value of environmental predictions, saving lives, protecting property, increasing economic vitality, protecting the environment, and informing policy choices. Indeed, forecasts in the subseasonal to seasonal (S2S) time range (defined in this report as 2 weeks to 12 months; see Box 1.1) have the potential to inform activities across a wide variety of sectors as many important decisions are made weeks to months in advance.

The potential of S2S forecasting has advanced substantially over recent decades, as improvements in numerical modeling, the Earth observing network, and understanding of sources of Earth system predictability in the so-called “gap” between short-range weather and climate timescales (see below) have enabled the development of extended-range weather and seasonal climate forecasts. As the availability and skill of seasonal climate forecasts—and more recently subseasonal predictions—has improved, S2S forecasts are increasingly being used in sectors such as agriculture, energy, and water resources management. However, there is enormous potential to further increase the benefits of S2S predictions. Many sectors have yet to exploit even the S2S information that is currently available. The user base could expand dramatically if the skill of S2S forecasts improves, more variables of the Earth system are explicitly forecast (e.g., a wider range of conditions of the ocean, cryosphere, and land surface), and users’ awareness of and ability to apply S2S information to important decisions and actions increases. Because so many critical planning and management decisions are made in the S2S time frame, it might be argued that the benefits of the longer range predictions have the opportunity to meet or exceed the current value of 0- to 14-day weather predictions if the quality, scope, and utilization of the forecasts can improve from their current state. S2S predictions may become even more valuable under anthropogenic climate change, because improved S2S forecasts could allow for the development of early warning systems that are becoming even more of a societal imperative in a warming world.

MOTIVATION FOR THIS STUDY

This report develops a vision for realizing the potential benefits of S2S Earth system predictions within the next decade. It identifies key strategies and proposes a research agenda with specific recommendations to guide progress toward that vision. There were four main motivations for initiating this study:

• The need to develop a research agenda to close the “gap” between efforts to improve numerical weather prediction (NWP) and climate modeling;

• The need to expand and improve S2S forecast capabilities beyond dynamical predictions of the atmosphere (i.e., to improve or develop S2S predictions of the oceans, land surface and cryosphere, as well as predictions of atmospheric variables such as aerosols);
• A desire to develop a more global S2S forecasting capacity, especially to meet needs related to national security and humanitarian response; and
• A changing computing environment that may open up both new opportunities and challenges for Earth system prediction.

As noted above and in the 2010 NRC report, Assessment of Intraseasonal to Interannual Climate Prediction and Predictability (hereafter NRC, 2010b; or ISI Report; see Box 1.2), S2S forecasting falls in a “gap” between the current modeling capabilities used for short- and medium-term prediction and those used in climate projections. Because of the short lead times involved with numerical weather prediction, efforts to improve weather forecasting have been focused on enhancing the accuracy of atmospheric and surface data for specifying initial conditions and on representing the short-term evolution of the atmosphere from this initial state. Earth system models that were first developed for making long-term climate projections have focused, in contrast, on representing Earth system processes that evolve more slowly (such as large-scale atmosphere and ocean circulation, the cryosphere, the state of land surface, and feedbacks between components) and how these processes are influenced by external climate drivers (e.g., greenhouse gas emissions, volcanic activity, other aerosols, and solar variability).

Although there is a traditional separation between research on weather and climate timescales, the boundaries between short-term and climate prediction are largely artificial (Shapiro et al., 2010). Because both fast- and slower evolving aspects of the climate system are important to conditions that develop in the 2-week to 12-month forecast range, S2S forecasting systems require close attention to initial conditions and high-fidelity representation of coupling and feedbacks between more slowly varying aspects of the Earth system. The potential to close this “gap” is now supported by a body of research indicating predictability in the Earth system at all timescales (e.g., Hoskins, 2013). In the S2S time range, this predictability arises in part from coupled ocean-atmospheric phenomena such as the El Niño–Southern Oscillation (ENSO) and the Madden Julian Oscillation (MJO), and in stratosphere-troposphere interactions associated with the Quasi-Biennial Oscillation (QBO) (see Box 1.3). Further S2S predictability may exist in other climate oscillations and their teleconnections, and in the Earth system response to slowly varying conditions in the ocean, land, and cryosphere, among other phenomena. Efforts are already under way in the United States and internationally to exploit these sources of S2S predictability, stretching the lead time of

weather timescale models forward and climate models backward, in part through the development of improved and more highly coupled Earth system models.

The continued development of coupled Earth system models also presents an opportunity to expand and improve S2S forecasts of environmental conditions well beyond the traditional weather variables, which represents a second major motivation for this report. There is a strong desire to develop more reliable S2S forecasts of conditions in the ocean and cryosphere and on the land surface, and meeting these needs is becoming more important as the financial and societal implications of managing environmental risk become more evident and larger in magnitude. Reliable ocean

forecasts on S2S timescales, for example, could improve the safety and effectiveness of commercial, military, and humanitarian operations at sea, in part by improving planning and ship routing by indicating ice-free and freeze-up likelihood as well as other ice and ocean eddy hazards. The desire for this type of S2S forecast highlights the importance of high-fidelity representation of ocean, sea ice, and land surface conditions in S2S forecast systems, in many cases for reasons beyond whether they feed back to influence the atmosphere.

A third major motivation for this report is the increasing desire for an enhanced forecasting capability globally. In particular, the Departments of Defense and State desire S2S forecasting capability that can best support U.S. engagement anywhere in the world. In addition, commerce, agriculture, and civilian hazard warnings currently at the national level could be expanded to cover more of the world. Developing a comprehensive and skillful global forecasting capability poses an additional challenge because, in many areas, only limited in situ weather data are publicly available for use in evaluating and improving forecasts.

Finally, accelerating computer and software capabilities could allow S2S prediction systems to operate with greater spatial and temporal resolution, more complete representation of interacting components of the Earth system, and more ensemble members for calculating uncertainties. Together with improved understanding of the physical process governing the Earth system’s dynamics and potential advances in the

ability to assimilate data into more sophisticated models, new computing capabilities could allow for significant gains in S2S predictions over the next decade.

Despite these needs and opportunities for enhanced Earth system forecasts in the S2S time range, a coordinated national research agenda aimed at strengthening the contributions of S2S forecasts to public and private activities has not yet emerged. For all of these reasons, the Heising-Simons Foundation, the National Aeronautics and Space Agency (NASA), and the Office of Naval Research (ONR) asked the National Academies of Sciences, Engineering, and Medicine to undertake a study aimed at outlining a 10-year research plan to advance the nation’s capacity to provide more skillful, comprehensive, and useful S2S predictions. The statement of task that guided the study (see Appendix A) asked the authoring committee to develop a strategy to accelerate progress on extending prediction skill for weather, ocean, and other Earth system forecasts from meso/synoptic scales to higher spatial resolutions and longer lead times, thereby increasing the nation’s research capability and supporting decision-making at medium and extended lead times.

In order to meet this request, the current study reviews present S2S forecasting capabilities and recommends a national research agenda to advance Earth system predictions at lead times of 2 weeks to 12 months. The study builds on previous reports that have described a grand vision to significantly advance forecasting accuracy, lead time, and prediction of nontraditional environmental variables (NRC, 1991b, 2008), as well as reports that have discussed opportunities and best practices for intraseasonal-to-interannual prediction (NRC, 2010b). In the years to come, the research agenda proposed here and the efforts that follow could produce increasingly accurate numerical models of the Earth system by describing its coupled interactions and future evolution, thus enhancing the value of weather, climate, and other Earth system forecasts to society.

THE REPORT ROADMAP

This report addresses the committee’s charge in seven subsequent chapters. Chapter 2 provides context for discussions in the remainder of the report by presenting an overview of the history and recent evolution of the field of S2S forecasting, descriptions of recent and ongoing research activities, and a summary of the current status and skill of operational S2S forecasting systems.

Chapter 3 covers decision-making contexts, applications for S2S forecasts, potential benefits of S2S predictions, attributes of effective forecasts, and user sensitivity to forecasting accuracy. The chapter also contains case studies, including western U.S. water

management, public health, and national security and defense, that provide more in-depth discussions of needs for and applications of S2S predictions.

Chapter 4 introduces sources of S2S predictability from natural modes of variability and teleconnections, as well as from the ocean, soil moisture, terrestrial snow, and sea ice and external forcing. The chapter includes recommendations to further predictability research in the S2S context.

Chapter 5 discusses in detail recent advances and activities needed to accelerate the improvement of S2S prediction systems, including discussions of gaps and research needs related to routine observations, data assimilation, and models, as well as calibration, combination, validation, and assessment of S2S forecast skill.

Chapter 6 covers research-to-operations in the context of current operational and research S2S prediction systems.

Chapter 7 presents findings and recommendations on infrastructure for computing, storage, programming models, shared software, and data cyberinfrastructure. The chapter also discusses institutional and workforce capacity building for S2S forecasting and decision support.

Chapter 8 concludes the report by presenting the committee’s vision for the future of S2S forecasts, as well as a summary of the committee’s proposed research strategies and agenda to advance S2S forecasting over the next decade.

This page intentionally left blank.